The new technology links wireless sensing with artificial intelligence (AI) to regulate when a patient uses an insulin pen or inhaler, and highlight potential errors in the patient's administration approach. "Some past work reports that up to 70% of patients do not take their insulin as prescribed, and many patients do not use inhalers properly," says Dina Katabi, the Andrew and Erna Viteri Professor at MIT, who is part of the research group that has developed the new AI solution. The group says that the system can be installed in a home and potentially help patients and caregivers by highlighting medication errors and reduce unnecessary hospital visits. 

"For example, insulin pens require priming to make sure there are no air bubbles inside. And after injection, you have to hold for 10 seconds," says Zhao on the official MIT blog about the announcement. "All those little steps are necessary to properly deliver the drug to its active site." At step, there is a possibility of error, especially when with no medical expertise at hand. 

The system can be classified into three main stages - first, a sensor traces a patient's activity in the 10-meter radius, using radio waves that reflect from their bodies. Next, AI checks for signs of a patient self-administering an inhaler or insulin pen. And finally, if it detects a flaw in the administration process, the system alters the patient or their health provider about the same. 

This innovation is an adaptation of a wireless sensing technology that the research group had previously utilised for monitoring people's sleeping positions which, through a radio-wave enabled wall-mounted device, would record movements by modulating signals and reflecting them to the device's sensor. "One nice thing about this system is that it doesn't require the patient to wear any sensors," says Zhao. "It can even work through occlusions, similar to how you can access your Wi-Fi when you're in a different room from your router."

Just like the sleep-monitoring device, the new sensor is stalled in the living space and uses AI to interpret the modulated radio waves. The researchers have trained the neural network to identify if insulin or an inhaler is being used. By reinforcement training, the device successfully identified 96% insulin uses and 99% inhaler uses! The device also proved to be useful in administering corrections. Since every step of administering insulin or an inhaler has precise steps in a sequence, the system can flag anomalies when it sees them. "By breaking it down into these steps, we can not only see how frequently the patient is using their device but also assess their administration technique to see how well they're doing," says Zhao.

The researchers say a key feature of their radio wave-based system is its noninvasiveness. "An alternative way to solve this problem is by installing cameras," says Zhao. "But using a wireless signal is much less intrusive. It doesn't show peoples' appearance."

The research was published in Nature Medicine in March 2021. The study was led by Mingmin Zhao, a PhD student in MIT's Computer Science and Artificial Intelligence Laboratory (CSAIL), and Kreshnik Hoti, a former visiting scientist at MIT and current faculty member at the University of Prishtina in Kosovo. Hao Wang, a former CSAIL postdoc and current faculty member at Rutgers University, Aniruddh Raghu, a CSAIL PhD student are the other co-authors on the study.